How to make a measurment graph flat to use as a Reference measurment?

[Maris]

Hello,
My point of interest is not acoustics, but measuring and comparing a rather large number of Large and Nearfield monitors as well a large number of Condenser, Dynamic and Ribbon microphones. For example I use a Genelec 1037B Large monitor for microphone frequency response measurements using Pink PN as a signal and RTA 1/3 octave meter. The problem I wish to solve is how to make one of the measurements made with my High Grade Measurement Mic as a Reference graph? How to set the graph flat and be able to measure all other microphones against this Flat Reference graph? I thought I could find this ability in the filter section, but I did not find such option. I got this idea about setting a measurement as a Reference plot from a professional microphone electronics designer who uses but an old measurement software exclusively because of this feature. He puts his own measurement mic in front of a generic speaker cabinet, makes a measurement, sets it as a Reference measurement which makes the graph flat and then takes my possibly damaged mic to see its plot against the Reference! Or he takes my Neumann KM184 mic pair, measures one of them, sets it as a Reference and then measures the other mic to see the deviation. Now I know if I can use them as a Stereo pair and also if these mics have retained their original performance properties.
Thanks for reading,
Maris

 

[Tony V.]

first of all welcome aboard AV NIRVANA!

no mic is going to give you a flat reading out of the box no matter how high end it is because every manufacturer designs the mic for a purpose (vocal, instrument or other) so it will have its own frequency curve. What you must have is a calibration file that gets inputted into REW and that file will give the mic your using a flat response. The challenge is you need to know the mic;s response beforehand to creat this file and most manufacturers dont give out that info.
Your best option is to get a UMIK-1 from miniDSP https://www.minidsp.com/products/acoustic-measurement/umik-1 it comes with the calibration file and works great with REW.

 

[Greg Dunn]

I would also point out that 1/3 octave isn't adequate to measure a sound source with if you're going to eventually use the mic to measure speakers (which in turn are listened to by human ears). The human ear perception/response aligns pretty well with 1/12 octave bandwidth; using broader filters leads to the possibility that you will miss peaks and dips which are audible to the ear.

 

[John Mulcahy]

Hi Maris, there are a couple of options:

You could save the measurement you want to use as a reference as text (File -> Export) then load that as a cal file in the measurements you want to compare it to (using the Change Cal... button).

You could use the trace arithmetic A/B function to divide the comparison measurement by the reference measurement, the result is the difference.

Note that in both cases the measurements should be at similar dB levels, if necessary you can use the offset controls (Add offset to data) to align them first. The results will sit at around 0 dB.

 

[Maris]

Thanks for the answers!
Tony, the mic I am using as a Reference mic is a custom built Omni measurement mic with a Panasonic 5mm capsule and custom electronics. The electronics are designed to serve not only for measurements, but for Pro recordings as well. It should have a reasonably flat graph. The large variety of my studio condenser mics have decently flat responses as well. They are Neumann KM-184, U-87, TLM-127, Schoeps ...
So I am not in need for a mic with a calibration file although that would be fine to have one. I trust in my own custom made mic. But thats not the problem because what I am doing is comparing this large variety of studio mics one against other. And the sound source is a studio monitor which obviously has a much more uneven frequency plot than any of the mics. Thats natural.So to be able to conveniently compare these different mic plots I wish I can set one of them as a flat Reference plot. The same goes for comparing different studio speakers using the same room, the same mic and setup.
So thank you John for your advice! I will try it out and I hope I will succeed.
And yes, I agree I would rather use 1/12 octave RTA. But for comparing the speakers I dont use Pink PN and RTA. I use the Measure feature of REW and this feature uses Sweep sine wave. I have even compared measuring a speaker with both the Measure function and the Generator with Pink PN and 1/3 RTA with Adjust levels box checked- the results were the same. Nevertheless I will stay with the standard Measure function for speaker measurements. Using Pink PN and the RTA for mic comparison gives some ease to this process.

Best wishes,
Maris

 

[Maris]

Hi Maris, there are a couple of options:

You could save the measurement you want to use as a reference as text (File -> Export) then load that as a cal file in the measurements you want to compare it to (using the Change Cal... button).

Note that in both cases the measurements should be at similar dB levels, if necessary you can use the offset controls (Add offset to data) to align them first. The results will sit at around 0 dB.

Thanks John!
The Change Cal option works. And now I know that I can use the Export Measurement to Text File to Load that as a Calibration File in the Preferences > Mic/SPL Meter
I will use this option only in Microphone plot measurements because I want to see how each of our 30 + mics sit against a Reference measurement Mic. This way all new measurements will be aligned against this Flat Reference plot. Otherwise I see how differently these mics have measured my Genelec speaker! I used the Measure function for our speaker measurements. But for some reason I prefered to use the Generator with Pink PN and RTA for my mic measurements. This way I could see the alignment of mics before saving to measurements. Now I see that if a measurement is saved from the RTA window it doesn't have the Change Cal button. And I guess that the Mic calibration file won't affect the Measurement captured and saved from RTA window. Am I right?
Otherwise I checked and got a conformation that the measurements made with the Measure function and Sine sweep were directly the same as made with Pink PN generator and RTA.

Best,
Maris

 

[John Mulcahy]

Cal files are applied to the RTA results, but their effect can't easily be separated from the underlying measurement data after it has been captured so there isn't an option to change the cal file for those.

 

[Maris]

Hello,
Thanks to John Mulcahy and his REW software I can now share my experience on how to test microphones. I have quite a lot of condenser, dynamic and ribbon mics in our studio and I wanted to know how they compare to each other. Here is one example.
We have some 10 Schoeps high end small diaphragm condenser mics of the same model. They are about 20 years in use so I wanted to check if they have retained their frequency response and sensitivity properties. My sound source obviously was Genelec 1037B studio monitor. I have a custom built trusted Measurement mic. I used the REW Measure function to do this. But the problem is that even a studio Pro monitor's frequency curve is far more uneven compared to the condenser mic's curve. So simply doing 10 measurements with Schoeps would give me curves hard to be compared.
John advised me the solution. I made a measurement of Genelec with my Measurement mic, exported this measurement to a text file and then in the Preference menu under the Mic tab loaded this text file as a Calibration for all subsequent measurements!
At first I made a new measurement with my measurement mic to see if it is now flat. And it really was. The same could be done in the measurement panel pressing the Change Cal button in the bottom of a measurement.
Now I made 10 measurements of all my Schoeps mics and what I got were 10 curves each showing their deviation from a flat frequency response. I also now see their sensitivity deviations! Now I can see which of the mics would need repairs or inspection! I think this is a reliable and efficient method.
Kindest Regards,
Maris

 

[Maris]

Hello,
One more question. When I open one of my mic measurement file the program loads all my 30 measurements. You can uncheck any of them. But most of time I want to compare only 2 measurements. Is there a way (a keyboard shortcut) to uncheck ALL measurements at once and then I could check the ones I need to look at?
Thanks,
Maris

 

[John Mulcahy]

Right click in the trace legend area.

 

[nospamatall]

Thanks for the answers!
Tony, the mic I am using as a Reference mic is a custom built Omni measurement mic with a Panasonic 5mm capsule and custom electronics. The electronics are designed to serve not only for measurements, but for Pro recordings as well. It should have a reasonably flat graph. The large variety of my studio condenser mics have decently flat responses as well. They are Neumann KM-184, U-87, TLM-127, Schoeps ...
So I am not in need for a mic with a calibration file although that would be fine to have one. I trust in my own custom made mic. But thats not the problem because what I am doing is comparing this large variety of studio mics one against other. And the sound source is a studio monitor which obviously has a much more uneven frequency plot than any of the mics. Thats natural.So to be able to conveniently compare these different mic plots I wish I can set one of them as a flat Reference plot. The same goes for comparing different studio speakers using the same room, the same mic and setup.
So thank you John for your advice! I will try it out and I hope I will succeed.
And yes, I agree I would rather use 1/12 octave RTA. But for comparing the speakers I dont use Pink PN and RTA. I use the Measure feature of REW and this feature uses Sweep sine wave. I have even compared measuring a speaker with both the Measure function and the Generator with Pink PN and 1/3 RTA with Adjust levels box checked- the results were the same. Nevertheless I will stay with the standard Measure function for speaker measurements. Using Pink PN and the RTA for mic comparison gives some ease to this process.

Best wishes,
Maris

If you're looking for reasonably flat microphone, you cannot choose cardioid, simply because of technology. They all show bass rolloff. In case of Neumann, don't use KM184, but KM183. See page 8 of user manual below